Backend Engineer, AI (Agent Systems)

Our Client is a AI start-up company. As a Backend Engineer, AI, you'll build and operate the backend infrastructure that powers every AI interaction within the product. You'll design and maintain production-grade systems that transform model capabilities into robust APIs and services consumed by mobile and desktop applications, with a strong focus on performance, observability, and operational excellence.

Responsibilities

• Build, deploy, and maintain backend services that power AI features in production.
• Design inference pipelines, orchesation layers, and service architectures around AI models.
• Develop scalable APIs that enable seamless integration between AI models, frontend applications, and internal services.
• Own production operations, including monitoring, logging, alerting, and incident response.
• Optimize inference performance through caching, batching, streaming, and other latency-reduction techniques.
• Improve system scalability, reliability, and cost efficiency based on production metrics and usage patterns.
• Collaborate closely with machine learning, product, and frontend teams to deliver high-quality AI-powered experiences.

Requirements

• Strong backend engineering experience building and operating production systems.
• Experience developing high-throughput, low-latency distributed services.
• Familiarity with AI inference architectures, including LLMs, embeddings, and multimodal models.
• Strong debugging and troubleshooting skills in distributed systems and production environments.
• Experience designing scalable APIs and backend services.
• A pragmatic, ownership-driven mindset with a focus on shipping reliable systems and continuously improving them through production insights.
• Technologies and tools: Python, Node.js, PyTorch, OpenAI, Anthropic, and open-source LLMs, SQL & NoSQL databases, Kubernetes, Docker

Success Looks Like

• AI backend services consistently deliver low-latency, high-throughput performance at production scale.
• APIs are reliable, well-designed, and enable seamless integration across frontend and machine learning systems.
• Production issues are detected early, resolved quickly, and followed by meaningful reliability improvements.
• Inference infrastructure is scalable, observable, and cost-efficient as usage grows.
• Continuous improvements driven by production data result in measurable gains in system performance, stability, and developer experience.